I am an IT geek physician. I have my an EHR which I created and control.
Today, I wanted to understand my diabetic practice a little more, so I dumped all my HbA1c data out of my EHR and into a spreadsheet where I was able to manipulate the data and learn a few things about my practice.
I learned that:
If my patient had a HbA1c ≥ 8, the likelihood that the HbA1c would be < 8 at the next visit is 68%.
If my patient had a HbA1c ≥ 8, the likelihood the HbA1c would be even higher at the subsequent visit is 29%.
If my patient had a HbA1c ≥ 8, the average change in the HbA1c at the next visit was -0.7.
If my patient had a HbA1c < 8, the likelihood that HbA1c at the subsequent visit would exceed 8% would be 15%.
Unfortunately, the other endocrinologists in my community cannot generate this type of information as they are dependent on the goodwill of the hospital IT people who must extract the data from the hospital’s EHR. While this may simply be a matter of asking for the data, in the real world, the physician may not know what they are looking for until after they “play” with the data.
I came to the realization that one of the big problems with the centralization of data in the large EHR silos is that we physicians have lost the ability, as individuals, to learn things about our practice from our data. We are dependent on the “generosity” of the health IT geeks, the administrators, our employers and those who own or control the EHR to give us the data. We are only allowed to see our data if it suits their priorities.
Just as importantly, we can not ask them to “give us our data” as we will not know ahead of time what we want to analyze. We will only know which data we need after we have had a chance to freely investigate all the data, as it is the “investigative” process that the relevant questions will be revealed.
This is most unfortunate for the future of healthcare and our patients.
Hayward Zwerling is a physician practicing in the Boston area.